As hardware for reproducing and storing high-resolution or high-quality video contents is being developed and supplied, there is a growing need for a video codec for effectively encoding or decoding high-resolution or high-quality video contents. Existing video codecs are designed to encode video on a basis of coding units having a tree structure according to a limited encoding method.
Image data in a spatial domain is transformed into coefficients in a frequency domain through frequency transformation. For fast frequency-transformation, a video codec divides an image into predetermined-sized blocks and performs discrete cosine transformation (DCT) on each of the blocks to encode frequency coefficients in units of the blocks. The existing compression systems perform block-based prediction to remove redundancies between color images. Existing compression systems generate parameters, which are to be used to perform video encoding and decoding, in units of pictures.